Fullscreen HTML5Med Res (??MB)HTML4

3D Printer Filament Drying Chamber This enclosure can store up to four 1kg reels of 3D printer filament, keeping them dry and ready for use at any time. You don’t even need to remove them – the filament can simply be fed to the printer through a small hole in its lid! Part 1 by Phil Prosser T he ability to produce functional 3D parts, either standalone or as part of a larger project, is incredibly useful. Over the last few years, 3D printer prices have fallen remarkably. You can now find some amazingly-priced 3D filament printers on the market. The major Australian electronics stores (Jaycar and Altronics) both stock “Creality” products, which I think are excellent. There are plenty of other good alternatives available online. My grandson, who wanted to buy printed parts, drew me into this. I pointed out that for the price of a handful of ‘bought bits’, we could buy our own 3D printer. So I did. I quickly found that being able to manufacture complex 3D parts was incredibly handy. Like most of these technical things, once you start, there is an amazing range of extras you might want or need. One surprising accessory is a filament dryer. It had not dawned on me that plastic filament can absorb moisture. However, PLA (polylactic Photo 1: the surface of the black boat is not smooth due to moisture in the filament. The white filament was dry, giving a much better result. 20 Silicon Chip acid), probably the most common filament these days, is sufficiently hygroscopic that moisture can become a real problem. 3D printers work by heating the plastic filament to around 200°C (or much hotter for materials like ABS) and extruding it through a small nozzle, typically 0.4mm in diameter. The printer acts like an X-Y plotter and deposits lines of melted filament where required, in layers, thus building the part. It is incredible to consider that a large print may have the printer laying down material in this manner for 12-24 hours, all without error. If that sounds too complicated to be reliable, well, you need to get many things right for the printer to work well. However, when set up correctly, reliable results can be achieved. I would say that most electronics hobbyists would have the inclination, skill and inquisitiveness to learn the tricks and tips required to keep a 3D printer running, but they certainly are not ‘set and forget’. When I first ran the printer, things went swimmingly well. However, I later realised that even a little moisture in the filament can cause problems when it is heated in the extruder. The moisture boils into steam, which pushes filament out of the extruder and causes ‘blobs’ on the print. Photo 1 tries to show the difference between fresh new filament (white) Australia's electronics magazine and some that had been lying around (black). All the printer knows is that it has driven the correct length of filament at the right time, but the ‘blobs’ mean it doesn’t end up exactly where it should be. So surfaces can get ‘blobby’, and you hear small popping noises during printing. While PLA certainly suffers from these problems, other materials, such as Nylon, also have a terrible reputation for being hygroscopic and hard to print with. While the printers themselves are competitively priced, I was not really into spending hundreds more on a fancy filament dryer. Some people use a food dehydrator, which, while cheap, does not handle multiple reels or allow you to feed straight from the dryer to your printer. I was convinced that I could easily make something to do the job with a handful of bits from the spares box, a leftover laptop power supply and maybe a microcontroller. We can even customise the size and shape to suit our workspace and needs. So, while we provide a complete parts list here, you can modify the design to reuse bits you already have, saving a few bucks. There does not seem to be a specific ‘right way’ to dry or, perhaps more correctly, dehydrate filament. All approaches use an elevated temperature and some form of timer. Some add air circulation, while a few incorporate siliconchip.com.au a mechanism to change the air in the box periodically. The idea of heating the filament in a sealed enclosure is that when the air in the enclosure gets hotter, it can hold a lot more moisture, so relatively speaking, the air is dryer. In other words, the relative humidity of the air in the box reduces as it is heated. Fig.1 shows that for a typical room at 20°C and 40% relative humidity (RH), there is about 6g of water per kilogram of air. If the box is sealed, there is always the same amount of water in the box. So, at 42°C, we see the relative humidity will be about 10%. Because the air is now quite dry (for its temperature), it pulls moisture from everything in the box. PLA filament that has absorbed moisture does not dry out quickly; drying times are typically 6-9 hours. By keeping the dryer sealed and including some desiccant, such as silica gel, in the enclosure, we can keep the filament dry and ready for use. If you will not use the dry filament for a while, it remains a good idea to seal it in a vacuum bag. and substantial protection circuitry. The second part is making an enclosure for the filament. There are several possible approaches, ranging from very simple to quite complicated. Choosing your approach to the container is probably the most critical choice, as the controller is not that complicated. We built two enclosures. The first was a custom one optimised for our needs and just a little bit fancy – see Photo 2 and the image above. The second was an 18L plastic tub into which we installed the controller and heater (Photo 3). The latter proved to be quick and simple to assemble and quite effective. It must be said that it looks a lot like a plastic tub, though. We will provide an overview of how to build the custom enclosure but will not go into great detail. If you are not confident in filling in the details yourself, stick to using the off-the-shelf plastic tub. Both enclosures use the exact same controller, but we have arranged the heating plates quite differently to suit the differing enclosure shapes. In both cases, we found that without adding insulation to the enclosure walls, we could achieve about 47°C inside with 50W of heating. Adding a layer of Corflute to the bottom and walls of the enclosures increased the temperature at that power level by well over 5°C, effectively reducing the amount of power needed to keep the enclosure at a given temperature. The unit is powered by either a 24V DC 4A plugpack or an 18-24V 3A+ DC laptop power supply. Is it just me who has a growing collection of these things, which seem to outlast the laptops they powered? Either way, it drives a resistive heater in the box via a control board, much of which is safety circuitry. We put a couple of small bags of silica gel in the box to absorb any The design This project has two distinct parts. The first is a filament dryer controller board. This is a standalone thermostat controller board that could equally be used to control an incubator or curing oven for painted parts. The board is essentially a thermostat with a timer siliconchip.com.au Fig.1: water in the air plotted against temperature for a range of different relative humidity (RH) values, from 10% to 90%. You can see how hotter air can contain a lot more moisture for the same RH figure. Australia's electronics magazine October 2024  21 moisture released by the filament and occasionally change the air in the box to expel excess moisture. Cat litter crystals are simply silica gel, so for $10 at the local supermarket, we got a huge bag of silica gel from which we make our own drying sachets. We just put it in paper envelopes to pop in the dryer. Our filament dryer hangs the reels on a rod and allows you to draw the filament straight from inside the dryer box. We decided to omit a fancy display, which technically is not hard but adds construction constraints and cost. During development, we noted that even with a fan circulating air in the dryer, the temperature throughout the box varied significantly. So, a temperature display may feel important, but it would only be indicative. Leaving out the display also avoids the need for a humidity sensor. This decision was hard but it keeps things simple and cheap. If the box is warm and you have fresh silica gel, after a couple of cycles, your filament will be as dry as it will get. Some really cheap humidity sensors are available online that you can pop in the box if you want to monitor it. Because we are making potentially combustible materials hot, we have taken a very conservative approach to the design to ensure that it is as safe as reasonably possible. Refer to the text box on safety analysis for a discussion of how key design drivers were arrived at. If you are designing your own enclosure, you should consider the hazards we list and satisfy yourself that your approach mitigates all hazards. The design presented here is mostly about implementing the control and safety systems identified in Table 1, which mandate the following inclusions: • A controller that maintains the Dryer in a safe state until the user deliberately starts a cycle. • A thermostat, allowing the temperature to be set from room temperature to 50°C. • A timer that allows a six- or ninehour drying period, then shuts the heater down. Table 1 – Hazard & Risk Assessment Hazard Initial Risk Mitigation Final risk High Implement a temperature control system. Limit the maximum energy available so the ultimate temperature without control is safe (50W gives a maximum of around 60°C). Low Short circuit or critical component failure Low Integrate thermal switches/fuses that disable the system at a safe temperature. Include a fuse in the design, to blow in case of a catastrophic short. Low Excessive heating since the control system does not sense the real temperature Moderate Include a fan to circulate air throughout the enclosure. Low Failure of fan results in loss of thermal control Low Integrate a ‘fan operating’ sensor and shut the heater down if the fan fails. Low Heating element contacts personnel Medium Mount heating resistors inside a plenum or behind sheet aluminium to minimise the likelihood of contact with personnel. Low Personal injury User touches energised part Medium Operate the dryer from an isolated plugpack with a low voltage output. Low Electric Shock Long-term heating results in auto-ignition of material Low A timer shuts the unit down after six or nine hours Low Fire and uncontrolled energy Enclosure operates unexpectedly Medium The system starts in an idle state. Force the user to press a start button to commence drying. Low Inadvertent operation Software fails Low Critical controls (thermal- and energy-related) are to be implemented in hardware. Low Inadvertent operation Heating element touching combustible material Medium Limit the heating power such that the element does not exceed 80°C. Mount the heating element so it is not in permanent contact with timber. Use polypropylene Corflute for insulation, which has an autoignition temperature of 288°C (flash point 260°C). Low Fire Misuse – user fills the enclosure with rags or paper Medium Integrate thermal cutout on heater plates at 90°C (high but safe). Low Fire Misuse – user covers the dryer with a blanket Medium Use a thermostat to control the internal temperature, with a safety shutdown & timer. Low Fire Uncontrolled heating, causing the enclosure to become excessively hot 22 Silicon Chip Australia's electronics magazine Consequence if not mitigated Damage or combustion of filament or enclosure siliconchip.com.au • Onboard fusing. • A thermal cutout on each heater element. • A thermal fuse on the controller board. • The maximum heating power is limited to 50W. • A ventilation fan that is integral to the controller board, ensuring airflow in the box. • An interlock that shuts down the heater if the ventilation fan stops. We have spread the heating across six 25W resistors, which dissipate 8W each into the large aluminium heating element. Even if everything fails, they will never get hot enough to create a hazard. We tested our two boxes with all controls disabled and determined that 50W of heating resulted in a maximum box temperature of no more than 60°C. Looking at what is on the market and having read a lot of tests on commercial filament dryers, most make wild claims as to the temperatures they achieve. We feel that 50-55°C is a good, safe temperature. If you want it to get hotter, you would need to increase the power or reduce the size of the box. The controller will accommodate that, but we advise you approach any changes with appropriate caution. You may have your own spin on how to build this; you could design a box that better suits your needs and use a surplus power supply. You could even reuse some different heating resistors. That will let you build a dryer for a fraction of the cost of a ‘bought one’, but make sure you follow our safety tips so everything goes well for you. We will first describe the controller and then present a couple of way it can be used. Photo 2: this DIY timber box can be sized to suit your needs. It has a rod for hanging the reels and convenient handles. The lid is removable and has a hole for feeding filament through. The controller The controller can operate from 18-24V DC, so you can recycle a laptop supply or similar power brick. It must deliver sufficient current for your resistor bank. The input is fused; select a fuse rating an amp or so above your expected maximum operating current. There is also a polarity protection diode that will dissipate about 2W; we have included heatsinking fills on the PCB, and this ‘extra power’ simply adds to the overall heating in the system. The controller is expected to be siliconchip.com.au Photo 3: this box from Bunnings doesn’t look as elegant and may be a little large for some people, but it’s much less work to prepare and does the job well. Australia's electronics magazine October 2024  23 installed inside the Filament Dryer, as that simplifies the wiring, and the temperature sensor is on the board. This means the controller will be operating at up to 50°C, perhaps a little more. That fine for most electronic components, but you will notice that we have specified high-temperature electrolytic capacitors and allowed for heatsinks on transistors Q1 and Q2. Circuit details The circuit is shown in Fig.2. An 8-bit PIC16F15214 operates as the timer, while an LM336-2.5 voltage reference (REF2) is used to produce a 2.5V reference, which is buffered by half of an LM358 op amp (IC1a). This is used in the temperature measurement circuit. The reason we have chosen the LM336-2.5 is it produces a reference voltage that is very stable over a wide temperature range. The LM336-2.5 has a variation of just 6mV over 0-70°C, so we can expect to see an error of less than a degree in temperature control over our operational range. The temperature sensor itself is a simple 1N4148 silicon diode (D6), using its -2.1mV/°C temperature coefficient. This is stable, reliable and used in many measurement circuits. The controller is a ‘Bang-Bang’ style, which simply turns the heating element on and off rather than implementing fancy control loops. This choice is again to keep things simple and cheap. The controller comprises half of the LM358 (IC1b), which compares the voltage across the sense diode to the temperature set voltage. We use the 2.5V reference voltage to set the current through the sense diode via a 4.7kW resistor. The same reference Fig.2: the circuit of the Filament Dryer Controller. REF2 and IC1a create a 2.5V reference (trimmed by VR1). This biases diode D6, the temperature sensor. The voltage across D6 and the setpoint from VR2/VR3 are compared by op amp IC1b to drive Mosfet Q2 for powering the heating elements. Microcontroller IC3’s timer limits the heating time and powers the fresh air fan periodically. 24 Silicon Chip Australia's electronics magazine siliconchip.com.au voltage generates the set voltage using trimpots VR1 and VR2 plus a couple of padder resistors. By using this very stable 2.5V reference, we can be assured that the current through the sense diode and the set voltage are constant over time and temperature. At room temperature, there is 400μA flowing through the sense diode, giving 0.56V across it. With the 12kW and 2.7kW padders and two 500W potentiometers, we get a temperature set point range of about 20-50°C. The reason we have included two pots is to allow us to use one (VR2) to set the minimum temperature to room temperature, while the other (VR3) is used to choose the temperature setpoint. With trimpot VR2 at the nominal value of 220W, the minimum voltage will be 0.489V (2.5V × 2.90kW ÷ [12kW + 2.92kW]). The maximum voltage will be 0.554V (2.5 × 3.42kW ÷ [18kW + 3.42kW]). The difference is 0.065V, and at 2.1mV/°C, that gives a spread of 31°C. Even using 1% resistors, the errors in the voltage divider are significant. If one is 1% high and the other is 1% low, the setpoint could move as much as 7°C. By adjusting VR2 so the minimum setpoint is room temperature, we can calibrate such errors out. The output of IC1b is low when the sensed temperature is below the setpoint and goes high when the temperature exceeds the setpoint. The 8.2MW resistor adds about 2°C of hysteresis by feeding back the output voltage to slightly shift the setpoint voltage. The ratio of the 8.2MW and 4.7kW resistors results in a shift of just a couple of milivolts, which is what we need. This stops IC1b from oscillating once the setpoint is reached. With the controller being flat out on or off, and the degree or two of hysteresis, the temperature control is not super precise. But for warming the filament to dry it out, that is OK. For the timer, we started by considering simple CMOS timer circuits and the venerable 555. To get a nine-hour period from these is not easy, so the cheapest way to make the timer was to use a PIC. These cost nearly $1.50 in single units, a fraction of the cost of the discrete solution, and can be programmed to do a huge range of jobs. We consider the timer to be an integral part of this design and strongly recommend against omitting it. Parts List – Filament Drying Chamber siliconchip.com.au Australia's electronics magazine 1 double-sided PCB coded 28110241, 126 × 93mm 1 18-24V DC 3A+ power supply (eg, laptop charger) 2 12V DC 40mm fans, 10mm-thick [Altronics F0010A] 1 40mm fan grille [Altronics F0012] 2 PCB-mounting M205 fuse clips (for F1) 1 5A 250V M205 fuse (F1) 1 77°C axial thermal fuse (F2) [Altronics S5631] 5 2-pin vertical polarised headers, 2.54mm pitch (CON1-2, CON4-5, CON7) [Altronics P5492] 5 2-pin polarised header plugs with pins [Altronics P5472 + 2 × P5470A each] 1 5-pin header, 2.54mm pitch (CON6; optional, for programming IC3 in-circuit) 1 PCB-mounting DC socket, 2.1mm ID or to suit power supply plug (CON8) 1 PCB-mounting 90° miniature SPDT toggle switch (S1) [Altronics S1320] 1 PCB-mounting 90° sub-miniature SPST pushbutton switch (S2) [Altronics S1498] 1 10kW side-adjust single-turn trimpot (VR1) 1 500W side-adjust single-turn trimpot (VR2) 1 500W 16mm single-gang linear potentiometer (VR3) 2 TO-220 micro-U heatsinks (optional) [Altronics H0627] 2 90°C normally-closed (NC) thermal switches/breakers (S3, S4) [Altronics S5612] Hardware (common to both versions) 1 3D-printed vent (“Vent Rotor.STL”, “Vent Rotor Base.STL” & “Vent No Fan.STL”) 1 3D-printed fan cover (“Fan Shroud.STL”) 6 M3 × 25mm panhead machine screws 18 M3 hex nuts & 32 M3 flat washers 1 3m length of high-temperature (90°C+) heavy-duty hookup wire 1 250mm length of 6mm diameter heatshrink tubing 1 2m length of 5-10mm wide open-cell foam adhesive tape 1 small tube of thermal paste Hardware (for plastic box version) 1 polypropylene box [Bunnings 0171464] 2 1.5mm-thick aluminium plates, 210 × 180mm Panhead machine screws: 8 M3 × 6mm, 32 M3 × 10mm, 8 M3 × 16mm, 6 M3 × 25mm Tapped spacers: 4 M3 × 15mm, 16 M3 × 25mm male/female hex spacers [Altronics H1243] Other: 58 M3 shakeproof washers, 46 M3 hex nuts Hardware (for timber box version) 2 3D-printed handles (“Filament Dryer Rail Tall.STL”) 1 sheet of 12mm MDF or plywood 1 1.5mm-thick aluminium plate, 330 × 225mm Panhead machine screws: 6 M3 × 6mm (30 if building lid), 16 M3 × 10mm, 4 M3 × 16mm, 24 M3 × 25mm, 1 M4 × 10mm (for attaching handle to lid) Tapped spacers: 12 M3 × 6mm (for lid), 10 M3 × 15mm Other: 42 M3 shakeproof washers, 38 M3 hex nuts Capacitors 1 470μF 35V 105°C electrolytic [Altronics R4865] 2 10μF 50V 105°C electrolytic [Altronics R4767] 7 100nF 50V multi-layer ceramic or MKT Semiconductors 1 LM358 dual single-supply op amp, DIP-8 (IC1) 1 LM336BZ-2.5 voltage reference diode, TO-92 (REF2) [Altronics Z0557] 1 PIC16F15214-I/P 8-bit microcontroller programmed with 2811024A.HEX, DIP-8 (IC3) 1 LM317T adjustable positive linear regulator, TO-220 (REG1) 1 BD139 80V 1.5A NPN transistor, TO-126 (Q1) 1 IRF540(N) 100V 30A N-channel Mosfet or similar, TO-220 (Q2) 2 BC548 30V 100mA NPN transistors, TO-92 (Q3, Q4) 1 BC338 25V 800mA NPN transistor, TO-92 (Q5) 1 BC558 30V 100mA PNP transistor, TO-92 (Q6) 4 1N4004 400V 1A diodes (D1, D3, D11, D13) 1 R250H or 6A10 400V 6A diode (D2) [Altronics Z0120A] 3 1N4148 75V 200mA diodes (D4-D6) 1 12V 0.4W or 1W zener diode (ZD10) 2 5mm red LEDs (LED7, LED8) 1 5mm green LED (LED12) Resistors (all ¼W 1% axial unless noted) 1 8.2MW 1 100kW 1 12kW 12 4.7kW 1 2.7kW 3 1kW 1 330W 1 47W 6 39W (18V), 47W (19-20V) or 68W (24V) 25W aluminium body resistors [Ohmite HS25 series] October 2024  25 Our dryer includes two fans. The first is to circulate air inside the box and it runs full-time. There is also a ventilation fan that runs briefly every 10 minutes. This is intended to draw fresh air into the box and to exhaust the hot (and possibly moist) air. This ventilation fan is driven by the PIC microcontroller. We do not want to continuously change the air in the enclosure, as it would require a lot of power to keep the temperature elevated. So our tiny PIC microcontroller drives the vent fan sparingly. Software The program in the timer is quite simple. At power-up, the PIC goes into an idle state, disabling the heater and ventilation. It stays in this state until the user presses the start button. This requires a deliberate action by the user. Once the start button is pressed, the timer moves into the running state. If IC3’s RA4 digital input is low, the timer drives its RA2 output low and counts nine hours. If RA5 is low instead, the output is low for six hours. After the selected time, the heater is switched off and the system goes back to the 26 Silicon Chip idle state. If the input is invalid, it remains idle. The PIC includes a secondary timer that drives digital output RA1 to switch on the ventilation fan every 10 minutes. The timer output and the output of the temperature sensor comparator are combined using open-collector transistors Q3 and Q4, which disable heater drive transistor Q2 when they are on. When the box is up to temperature, the output of IC1b goes high, switching on Q3, which disables the heater. Green LED12 is in series with this output, and lights showing that the set temperature has been achieved. Switching the load on is implemented using an IRF540 or similar power Mosfet with a gate pullup resistor to 12V. The gate drive pullup is derived from the ventilation fan power supply, which might seem an odd choice. The ventilation fan draws current through D11, D13 and the parallel 47W resistor. The specified fan draws 60mA in operation and develops 1.2V across these diodes. This voltage switches on Q6 on via its 4.7kW base resistor, which forms the Mosfet gate drive. If the fan stalls, its internal controller reduces its supply current to 2mA and attempts to restart it every few seconds. This 2mA current only generates 94mV across the 47W resistor, which is not enough to switch Q6 on, and consequently the Mosfet gate drive is removed. Thus, we disable the heater if the ventilation fan is stalled or not working. For Q2, pretty much any TO-220 package, low-RDS(ON) N-channel Mosfet will work. They virtually all have the same pinout. If you want to use a different Mosfet from our recommended part, look for one with an RDS(ON) under 0.1W. For example, the MTP3055V has an RDS(ON) of 0.18W and for a load current of 3A, it will dissipate 1.6W (3A2 × 0.18W). That would demand the use of a flag heatsink; there is room for this on the PCB. The recommended IRF540 has an RDS(ON) of 0.077W and will dissipate 0.7W at 3A (or 0.4W for the IRF540N version), which will make it warm but it won’t require a heatsink. Photo 4: the top side of the prototype PCB. The fan is mounted to the underside using four M3 x 16mm machine screws with matching hex nuts. Australia's electronics magazine siliconchip.com.au There are two headers for wiring up the heater resistors. This allows you to run separate wiring to two banks of resistors, making the wiring and layout easier in some builds. The current rating of the recommended Altronics P5492 headers is 3A, so you could get away with using just one. We have included a thermal fuse in the power supply to the Mosfet. The specified fuse has a current rating of 10A AC but in our application, we are breaking nominally 2A DC. The fuse does not have a DC rating but that is well within its capacity. This device will fuse at 77°C, and will hold at 55°C continuously. Should your enclosure exceed 55°C for extended periods, you may trigger this protection. The heater We performed a number of tests on the boxes we’re presenting and determined that we need 50W to heat our enclosures to 50°C reliably in a 20°C room. This is also a good maximum, as per the safety considerations we touched on earlier. To allow us to spread the power around the enclosure, we are using six 25W resistors mounted to large heatsinks. We have used 68W devices, which at 24V will dissipate 50W in total. To spread the heat, we used a 330 × 225mm aluminium sheet folded to fit inside our timber box, or two 210 × 180mm panels for the plastic box. If using a 19V supply, the heating resistor values need to be reduced to 47W to keep that 50W target. We recommend the cheapest aluminium case power resistor we could find, mentioned in the parts list. The cost is around $20 for six, so you can save some decent money reusing parts you have. It is important that the devices you select can be bolted to the heat spreader, as this ensures they do not get hot enough to create a hazard. We tried using 10W ceramic resistors each dissipating 5W. While they were operating within their specification, their surface temperature of over 130°C would have the potential to create a hazard if combustible material fell onto them. Safety considerations for the Filament Dryer In designing the controller, we undertook a hazard assessment and developed controls for each hazard we identified, seeking to mitigate these hazards as much as reasonably practicable. This in broader engineering often forms part of a “Safety Engineering Program”. This process involves identifying credible hazards them applying the ‘hierarchy of controls’ which, in order, are: ● Eliminate the hazard ● Substitute to avoid / minimise the hazard ● Apply engineering controls ● Add administrative controls (how it is used) ● Use Personal Protective Equipment (PPE) In safety engineering, there is an important differentiation between a hazard, which is a potential outcome, and the risk this represents, which considers the likelihood of this occurring. The intention of applying the hierarchy of controls is to mitigate and minimise the overall risk of a system. Our hazard assessment was undertaken to inform the design of the project and to shape the solution, both to minimise the underlying hazards in the design and also to apply substitutions, engineering and administrative controls to further mitigate residual risks. By keeping a record of the approaches to managing safety, and building those into the design, we can then test the project to ensure that these controls do what we expect. The hazards and controls we identified for the filament dryer are shown in Table 1 (Hazard & Risk Assessment). Some significant changes in design were implemented. Those practised in the safety art will note that we have picked parts of a larger process to document here, as a full safety program is comprehensive and at times less than fascinating. We have, however, included some important elements for your consideration when making your own version of this. Next month The second and final article next month will have the construction and testing details, including building or adapting and then insulating the container. SC Photo 5: the Filament Dryer in use, showing how filament is drawn from the container. siliconchip.com.au Australia's electronics magazine October 2024  27